Task Success Metrics (Real, WM, W2R, CR) on TextWorld
100RealGPT-4-turbo
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-4-turboWorld Model=Qwen2.5-7B WorldModel2025.12 | 100 | 99.42 | 98.84 | 99 | |
| GPT-4.1World Model=Qwen2.5-7B WorldModel2025.12 | 100 | 100 | 100 | 100 | |
| GPT-5World Model=Qwen2.5-7B WorldModel2025.12 | 100 | 100 | 100 | 100 | |
| Gemini-2.5-flashWorld Model=Qwen2.5-7B WorldModel2025.12 | 100 | 100 | 76.3 | 76 | |
| Claude-sonnet-4.5World Model=Qwen2.5-7B WorldModel2025.12 | 100 | 100 | 100 | 100 | |
| GPT-4-turboWorld Model=Llama3.1-8B WorldModel2025.12 | 100 | 97.69 | 93.06 | 93 | |
| GPT-4.1World Model=Llama3.1-8B WorldModel2025.12 | 100 | 98.27 | 94.22 | 94 | |
| GPT-5World Model=Llama3.1-8B WorldModel2025.12 | 100 | 98.84 | 94.8 | 95 | |
| Gemini-2.5-flashWorld Model=Llama3.1-8B WorldModel2025.12 | 100 | 99.42 | 93.06 | 93 | |
| Claude-sonnet-4.5World Model=Llama3.1-8B WorldModel2025.12 | 100 | 93.33 | 90 | 90 | |
| GPT-4oWorld Model=Qwen2.5-7B WorldModel2025.12 | 98.84 | 100 | 96.53 | 98 | |
| GPT-4oWorld Model=Llama3.1-8B WorldModel2025.12 | 98.84 | 97.11 | 90.17 | 91 | |
| GPT-4o-miniWorld Model=Qwen2.5-7B WorldModel2025.12 | 97.44 | 100 | 69.36 | 71 | |
| GPT-4o-miniWorld Model=Llama3.1-8B WorldModel2025.12 | 97.44 | 92.48 | 57.8 | 59 |